EP2591338A1

EP2591338A1 - Apparatus and method for dating a body or body sample

Info

Publication number: EP2591338A1
Application number: EP11738425.5A
Authority: EP
Inventors: Maurice Aalders; Rolf Bremmer
Original assignee: Academisch Medisch Centrum
Current assignee: Academisch Medisch Centrum
Priority date: 2010-07-09
Filing date: 2011-07-11
Publication date: 2013-05-15
Also published as: GB201011585D0; US9029161B2; WO2012004417A1; US20130196443A1

Abstract

A method and apparatus for dating a dating a body sample, such as blood, includes taking at least one spectroscopic measurement (37) of the sample at least two predetermined positions in the spectrum having spectral characteristics corresponding to at least two predetermined substances present in the sample that have a time varying relationship with each other. A measured relative concentration of each of the predetermined substances is then determined (41) from the measurement, and the measured relative concentrations of the two predetermined substances is compared (43) with a known variation of the relative concentrations of the two predetermined substances over time. A good fit of the measured relative concentrations to the known variation of the relative concentrations is then determined, so as to provide an indication of the age of the sample (44). Alternatively, instead of measuring the relative concentrations of each of the predetermined substances, the rate of change of the relative concentrations is determined.

Description

11 061785

- 1 -

APPARATUS AND METHOD FOR DATING A BODY OR BODY SAMPLE

Field of the Invention

The present invention relates to an apparatus and method for dating a body or body sample, preferably, but not exclusively, a sample of a body fluid, such as blood, which is external to the body, for example, having caused stains on an external substrate, such as a floor, wall, and clothing.

Background of the Invention

It is well known that dating of external blood stains is a useful forensic technique to determine when a particular injury was caused to a person. Thus, the injury may have been such that the skin was punctured and blood was spilt. Alternatively, traces of other bodily fluids, such as saliva may be found. In both cases, however, there is a need to accurately determine the age of the stain or sample, so as to determine when the body fluid was released from the body.

In general, determination of the age of externa! blood stains has been carried out by a variety of methods. Although, historically, this was carried out by very roughly estimating the age based on the colour of the blood stain, more "scientific" methods have recently been employed. Most such methods rely on the fact that, when blood leaves the human body, haemoglobin in the blood is oxidised to met-haemoglobin. Various methods have therefore tried to use different characteristics of haemoglobin and met-haemoglobin to try to determine their relative concentrations and, from those, to try to determine the age of the blood stain. However, some of these methods require very complicated and advanced technical equipment, so that they cannot easily be done at the scene of the injury, for example a crime scene. In other cases, known techniques often compromise the blood traces by using chemicals or by requiring that the stains be taken to a laboratory in order to properly analyse them. Still others of the methods have proved not to be reliable since they provide too high a deviation compared to the actual age of the stains. One recent new technique that has been suggested is to use atomic force microscopy for high-resolution imaging of erythrocytes in a blood sample to detect elasticity changes on a nanometer scale (see the paper "Age 61785

determination of blood spots in forensic medicine by force spectroscopy" by Stefan Strasser, Albert Zink, Gerald Kada, Peter Hinterdorfer, Oliver Peschel, Wolfgang M. Heckl, Andreas G. Nerlich and Stefan Thalhammer published in Forensic Science International, Volume 170, Issue 1 , 20 July 2007, Pages 8- 14).

In PCT Patent Application No. WO 2009/130580 there is described a technique for dating blood samples by taking a series of spectroscopic measurements of the sample to determine the concentrations of oxy- haemoglobin and met-haemoglobin at several spaced points in time. The ratios of the concentrations at the different points in time are then analysed to estimate when their concentrations were at a limit of their concentrations, thereby providing an indication of the age of the blood sample.

Although the above method is an improvement on the previous techniques, the estimate is based on the assumption that the reaction rate prior to the first measurement being made is the same (or at least roughly the same over the period) as the reaction rate over the time when the measurements are taken. Since it requires a series of measurements to be made, it can take some time for the estimate of the age of the sample to be determined. More accurate methods are therefore, of course, desirable.

Apart from body samples, it will be apparent that the assumption that the reaction rate prior to the first measurement being made is the same as the reaction rate when the measurements are taken also applies to other methods of dating of both body samples, and, in some cases, to methods of dating bodies, for example by taking the temperature of a body either once or several times and then assuming that it is cooling at a particular rate. However, such methods can be inaccurate because the cooling rate may depend on environmental factors.

Brief Summary of the Invention

The present invention therefore seeks to provide an apparatus and method for dating a body sample, which overcomes, or at least reduces the above-mentioned problems of the prior art.

Accordingly, in a first aspect, the invention provides a method of dating a body sample comprising taking at least one spectroscopic measurement of the sample, the measurement including at least two predetermined positions in the spectrum, the positions having spectral characteristics corresponding to at least two predetermined substances present in the sample that have a time varying relationship with each other, determining a measured relative concentration of each of the predetermined substances present in the sample from the spectroscopic measurement, comparing the measured relative concentrations of the at least two predetermined substances with a known variation of the relative concentrations of the at least two predetermined substances over time, and determining a good fit of the measured relative concentrations to the known variation of the relative concentrations, so as to provide an indication of the age of the sample.

According to a second aspect, the invention provides a method of dating a body sample, comprising taking at least two spectroscopic measurements of the sample at different times, the measurements including at least two predetermined positions in the spectrum, the positions having spectral characteristics corresponding to at least two predetermined substances present in the sample that have a time varying relationship with each other, determining a measurement of change in the relative concentration of each of the predetermined substances present in the sample from the spectroscopic measurements, comparing the measurements of change in the relative concentration of each of the predetermined substances with a known variation of the changes in the relative concentrations of the predetermined substances over time, and determining a good fit of the measurements of changes in the relative concentration to the known variation of the changes in the relative concentrations, so as to provide an indication of the age of the sample.

In a preferred embodiment, both methods can be used, if appropriate, for example, if the changes in the relative concentrations of each of the predetermined substances are small.

In one embodiment, the sample is a body fluid, which may be blood. Preferably, the predetermined substances comprise two or more of oxy- haemoglobin, met-haemoglobin and haemichrome.

The measurement(s) preferably includes at least three predetermined positions in the spectrum, the positions having spectra! characteristics corresponding to at least three predetermined substances, the three P2011/061785

predetermined substances comprising oxy-haemoglobin, met-haemoglobin and haemichrome.

Preferably, the spectroscopic measurement(s) comprise reflectance, Raman or fluorescence spectroscopic measurement(s).

According to a third aspect, the invention provides a method of dating a body or body sample, the method comprising measuring one or more time- varying parameters of the body or body sample, the time-varying parameter(s) having a rate of time variance that changes according to at least one environmental factor, providing a database of rates of time variance of the time- varying parameter(s) for the at least one environmental factor, estimating the at least one environmental factor for the particular environment(s) where the body or body sample was located prior to the at least one measurement being made, and determining an estimate of the age of the body or body sample by utilising the at least one measurement of the or each time-varying parameter and the rate of time variance for that time-varying parameter for the estimated at least one environmental factor.

In one embodiment, the time-varying parameter comprises temperature. Preferably, the time-varying parameter(s) comprises two or more of:

concentration of oxy-haemoglobin;

concentration of met-haemoglobin; and

concentration of haemichrome.

The environmental factor preferably comprise one or more taken from: ambient temperature;

ambient humidity;

thermal conductivity of a substrate on which the body or body sample is located; and

absorption by a substrate on which the body or body sample is located.

According to a further aspect, the invention provides an apparatus for dating a body sample comprising a device for taking at least one spectroscopic measurement of the sample, the measurement including at least two predetermined positions in the spectrum, the positions having spectral characteristics corresponding to at least two predetermined substances present in the sample that have a time varying relationship with each other, and a processing device for determining a measured relative concentration of each of the predetermined substances present in the sample from the spectroscopic measurement, comparing the measured relative concentrations of the at least two predetermined substances with a known variation of the relative concentrations of the at least two predetermined substances over time, and determining a good fit of the measured relative concentrations to the known variation of the relative concentrations so as to provide an indication of the age of the sample.

According to still further aspect, the invention provides an apparatus for dating a body sample comprising a device for taking at least two spectroscopic measurements of the sample at different times, the measurements including at least two predetermined positions in the spectrum, the positions having spectral characteristics corresponding to at least two predetermined substances present in the sample that have a time varying relationship with each other, and a processing device for determining a measurement of change in the relative concentration of each of the predetermined substances present in the sample from the spectroscopic measurements, comparing the measurements of change in the relative concentration of each of the predetermined substances with a known variation of the changes in the relative concentrations of the predetermined substances over time, and determining a good fit of the measurements of changes in the relative concentration to the known variation of the changes in the relative concentrations so as to provide an indication of the age of the sample.

In a preferred embodiment, if the changes in the relative concentrations of each of the predetermined substances are small, the processing device further determining a measured relative concentration of each of the predetermined substances present in the sample from the spectroscopic measurement, comparing the measured relative concentrations of the at least two predetermined substances with a known variation of the relative concentrations of the at least two predetermined substances over time, and determining a good fit of the measured relative concentrations to the known variation of the relative concentrations so as to provide an indication of the age of the sample.

In one embodiment, the sample is a body fluid, which may be blood. Preferably, the predetermined substances comprise two or more of oxyhemoglobin, met-haemoglobin and haemichrome.

The measurement(s) preferably includes at least three predetermined positions in the spectrum, the positions having spectral characteristics corresponding to at least three predetermined substances, the three predetermined substances comprising oxy-haemogiobin, met-haemogiobin and haemichrome.

More preferably, the reflectance spectroscopic measurements include near infra-red wavelength radiation of above 850nm or even more preferably between 1150nm and 2500nm.

In a further aspect, the invention provides an apparatus for dating a body or body sample, comprising a device for taking at least one measurement of one or more time-varying parameters of the body or body sample, the time- varying parameter(s) having a rate of time variance that changes according to at least one environmental factor; and a processing device for obtaining rates of time variance of the time-varying parameter(s) for the at least one environmental factor, for the particular environment(s) where the body or body sample was located prior to the at least one measurement being made, from a database of different rates of time variance of the time-varying parameter(s) for different environments, and determining an estimate of the age of the body or body sample by utilising the at least one measurement of the or each time- varying parameter and the rate of time variance for that time-varying parameter.

In one embodiment, the time-varying parameter comprises temperature.

Preferably, the time-varying parameter(s) comprises two or more of: concentration of oxy-haemoglobin;

concentration of met-haemoglobin; and

concentration of haemichrome.

The environmental factor preferably comprises one or more taken from: ambient temperature;

ambient humidity;

thermal conductivity of a substrate on which the body or body sample is located; and EP2011/061785

absorption by a substrate on which the body or body sample is located.

Brief Description of the Drawings

Several embodiments of the invention will now be more fully described, by way of example, with reference to the drawings, of which:

Figure 1 is a schematic diagram of an apparatus constituting a first embodiment of the invention;

Figure 2 is a schematic diagram of an apparatus constituting a second embodiment of the invention;

Figure 3 is a schematic diagram of an apparatus constituting a third embodiment of the invention;

Figure 4 shows absorption spectra for haemoglobin (Hb), met- haemoglobin (metHb) and oxy-haemoglobin (Hb0₂);

Figure 5 is a schematic representation of haemoglobin kinetics both inside the body and external to the body;

Figure 6 shows absorption spectra for met-haemoglobin, oxy- haemoglobin and haemichrome (HC);

Figure 7 is a schematic flowchart of a first process for determining the time at which blood was first exposed to air;

Figure 8 is a schematic diagram of the geometry of a simple model of light reflectance from two layers;

Figures 9a and 9b are schematic graphs showing the linear fit of the three components on a measured diffuse reflectance signal for a bloodstain that is ¼ and 63 days old, respectively;

Figure 10a is a graph showing the relative concentrations of Hb0₂, metHb and HC in a bloodstain over time up to around 60 days;

Figure 10b is a graph showing the relative concentrations of Hb0₂, metHb and HC in a bloodstain over time up to around 500 days;

Figure 11 is a graph showing the age estimation for test samples to compare the estimated age with the known actual age;

Figure 12 is a schematic flowchart of a second process for determining the time at which blood was first exposed to air; Figure 13 is a graph showing experimental data of how the relative concentration of Hb0₂ in a bloodstain changes over time at different temperatures;

Figure 14 is a line graph generated from the graph of Figure 13 showing how the relative concentration of Hb0₂ in a bloodstain changes over time at different temperatures;

Figure 15 is a graph showing how the relative concentrations of Hb0₂ in a bloodstain changes over time at different humidities;

Figure 16 is a table showing different values for equation parameters equivalent to the lines in the graph of Figure 14;

Figure 17 is a graph showing experimental data of how the relative concentration of HC in a bloodstain changes over time at different temperatures;

Figure 18 is a graph of white and coloured cotton t-shirts background spectra in the near infrared part of the spectrum;

Figure 19 is a graph of first derivative spectra of blood stains of 5 hours and 77 days old; and

Figure 20 is an age estimation of back cotton using the predetermined regions in the near infrared part of the spectrum.

Detailed Description of the Drawings

Referring first to Figure 1 , there is shown a schematic apparatus for measuring the age of bloodstains on a substrate. As shown, the bloodstain 1 is present on a host material 2. The bloodstain 1 is illuminated by light from a light source 3 via a first optical fibre 4. The light is directed onto the bloodstain from the end of the first optical fibre 4 via a linear polarizer 5. A second optical fibre 6 is arranged next to first optical fibre 4 to detect light that is reflected from the bloodstain 1 and which passes back through the linear polarizer 5. Such reflected light is, of course, low-order scattered, since its polarization has not been substantially changed. This low-order scattered reflected light is then transmitted via the second optical fibre 6 to a spectrograph 7. The results of the spectrograph, which will be further explained below, are then passed to a computer 8 for processing. A cross polarizer 9 is provided next to the linear polarizer and a third optical fibre 10 is arranged to detect light reflected from the bloodstain 1 that passes through the cross polarizer. This reflected light will be 61785

- 9 - high-order scattered light and this light is transferred via the third optical fibre 10 to a second channel of the spectrograph. This light will contain information of the host material, later to be used for correction. Low-order scattered light comes from the top of the blood and high-order from the bottom (or from the host material).

Figure 2 shows a second embodiment of similar apparatus to that of Figure 1 , but with different forms of the same basic elements. Thus, in this case, the bloodstain 1 is again on the host material 2, but this time it is illuminated by light produced by several Light Emitting Diodes (LEDs) 11 controlled and driven from an LED driver 12. The LEDs 11 are different, so that each emits light of a different wavelength or waveband from that of the other LEDs 11. The light from the LEDs 1 illuminates the bloodstain 1 via a first lens 13 and a linear polarizer 14. In this case, a second lens 15 directs the light reflected from the bloodstain and transmitted back through the linear polarizer 14, i.e. the low-scattered light, to several photo detectors 16, each of which has a different filter 17 between it and the second lens 15, so that each photodetector 16 detects reflected light of a different wavelength or waveband from that of the other photodetectors 16. The filters 17 are matched to transmit light of the wavelengths or wavebands that are emitted by the LEDs 11. The outputs from the photodetectors 16 are passed to a signal processing unit 18, which is coupled to a computer 19. The computer 19 is also used to control the LED driver 12. This implementation can be easily designed into a hand held measurement device, to enable ease of use at, for example, a crime scene. Again, low-order scattered light comes from the top of the blood.

Figure 3 shows a third embodiment of the apparatus. In this case, the bloodstain 1 is illustrated as being in a vertical plane, for example on a wall, !n this case, a hand-held apparatus 20 includes a light guide 21 having an appropriate optical system at an end 22 thereof for transmitting illuminating light from an input light module 23 through the optical system and out onto the bloodstain. Reflected light is then detected through the optical system and transmitted through the light guide 21 to a Spectral Imaging Camera 24. The Spectral Imaging Camera 24 includes both a colour camera 25 and an intensified camera 26. The detected light from the bloodstain 1 transmitted through the light guide 21 passes through a Liquid Crystal Tunable Filter (LCTF) 27 from where it is directed to either the colour camera 25 or the intensified camera 26 (or both). The LCTF operates as a tunable band pass filter, the centre wavelength of which can be rapidly tuned across a wide spectral range. The Spectral Imaging Camera 24 is coupled to a computer 28, which both receives the output data from the Spectral Imaging Camera 24 for analysis and controls the Spectral Imaging Camera 24, including the colour camera 25, the intensified camera 26, and the LCTF 27 via LCTF control electronics 29. Advantages of the LCTF, apart from electronic tuning, include high spectral resolution, simple to integrate in imageing systems and fast response time. Any filtering wavelength in the range from 400 to 720 nm can be rapidly selected either by random access or sequential tuning which, in the last case, yields a full reflection spectrum for every pixel on the CCD (= every position on the tissue). In addition, both filtered (spectral) and unfiltered (white light) images are available to facilitate the direct association of spectroscopic features with specific anatomical sites.

The most direct mode of spectral imaging is to illuminate the target area while acquiring a series of images assessing the remittance spectrum of every pixel. This data set is called a 'spectral cube', containing information about the absorbers in the tissue which allows determining relative concentrations of the absorbers. One such LCTF is manufactured by Cri (UK) and has an active diameter of 20 or 35 mm, a tuning range from 400 to 720 nm (which means that a separate detector is required for measuring water absorption), a throughput of less than 50% and a filtering bandwidth of nominally 7 nm (collimated light).

However, it will be appreciated that other filters can be used depending on the application and the accuracy of the results required. Thus, for example, an Acousto-Optical tuneable filter (AOTF), could be used instead of the LCTF. In any event, both types of filter produce data in the form of a spectral cube, which contains information about the absorbers in the tissue which can be used to determine relative concentrations of the chromophores. The reflected light at each position will be used to calculate the corresponding absorption spectrum at each pixel or group of pixels. This will provide the position dependent concentrations of the chromophores.

Inside a healthy human body, haemoglobin molecules are mainly present in two forms: one without oxygen: de-oxyhaemoglobin (Hb) and one saturated T EP2011/061785

- 1 1 - with oxygen; oxy-haemoglobin (Hb0₂). The average saturation level of arterial blood is >90% and of venous blood is >70%. In addition, only a small part {~1 %) of Hb0₂ is auto-oxidized into a third form, met-haemog!obin (met-Hb) . Reductase protein cytochrome b5 will then reduce met-Hb back to Hb.

It has been found, however, that outside the body, blood will first totally saturate to Hb0₂ as soon as it comes in contact with the oxygen in the atmosphere. This autoxidation process will continue and due to a decreasing availability of cytochrome b5, necessary for reduction of met-Hb, the transition of Hb0₂ into met-Hb will become irreversible. Once the haemoglobin molecules are autooxidized to met-Hb they will denature to haemichrome (HC). HC is formed through changes of protein conformation so that atoms endogenous to the protein become bound to the iron at the sixth ligand. A schematic overview of the haemoglobin kinetics, of both in-vivo and extracorporeal blood is shown in Figure 5.

Figure 4 shows absorption spectra for haemoglobin (30), met- haemoglobin (31) and oxy-haemoglobin (32). It will be seen that haemoglobin has an absorption peak at 556 nm, oxy-haemoglobin has a pair of peaks at 542 and 578 nm, and met-haemoglobin has a peak at 630 nm. Figure 6 shows the absorption spectra for Hb0₂, met-Hb, and HC. It will be seen that they differ at over 450 - 800 nm. Thus, by using Diffuse Reflectance Spectroscopy at these wavelengths to monitor this chemical process in bloodstains, a more accurate determination of the age of the bloodstain can be determined.

It will thus be seen that by using that by measuring the detected light, one can determine the (relative) quantities of the contents of the measured volume. From the total reflectance spectrum for the bloodstain, for the host material, for example coloured cloth, and a reference spectrum for the light source, one can correct for the properties of the light source and obtain the corrected reflectance of the bloodstain. From these spectra, the relative concentrations of the chosen substances in the bloodstain can be determined.

Thus, the process of determining the relative quantities of oxy- haemoglobin, met-haemoglobin, haemichrome and water in order to calculate the moment of deposition of extracorporeal blood is based on a spectroscopic measurement technique which only measures the low order scattered light and fits the resulting data using a procedure that allows for the elimination of P2011/061785

- 12 - background artefacts and scaling of the absorption curves. The process is shown generally in Figure 7.

As can be seen, the process starts by warming up and stabiiising the light source 34. The time at which a stable output is reached depends on the type of light source. A halogen lamp requires 10 - 30 minutes, but LEDs can be used much sooner, although they have to be kept at a constant temperature. Once the light source is stabilised, in order to be able to correct for the spectrum of the particular light source being used, a reference spectrum is measured (35) using a neutral reflector, such as a spectralon. In this case, the light from the light source is made incident onto the spectralon and the reflected light is passed to the detector using the same light path, for example optical fibres if they are being used, so that a spectrum including the light source, the light paths and the detector characteristics can be obtained as a reference. Next, the reflection spectrum from the host material is obtained (36). Clearly, this will bear the characteristics of the reference spectrum plus the absorption and reflection characteristics of the host material. After these two references are obtained, measurement on the bloodstain is commenced (37). If desired, one or more reflectance spectra may be measured at different positions on the bloodstain (38).

In essence, the absorbance of the host material and of the bloodstain can be found from:

From the measurements of the spectra, the absorbances of the host material and the bloodstain can be calculated (39).

This is illustrated schematically in Figure 8, where the geometry of the situation, is shown in simplified form considering a simple model of two layers, being the blood stain 1 and the host material 2. As will be apparent, the intensity of the light detected at the detector iood, whether it be the spectrograph 7, the photodetector 16 or the Spectral Imaging Camera 24, at a particular wavelength, will depend on the intensity of the incident light lo directed onto the bloodstain and will depend on the amount of light reflected by the bloodstain, reflected by the host material and then transmitted by the bloodstain. This can be expressed, simply as:

J T

¹ blood — ¹ 0

where

Ibiood is the intensity of the received light;

io is the intensity of the incident light;

fJa,host(A) is the absorption coefficient of the host material {in mm^"1) at wavelength λ; and

Va.bift) is the absorption coefficient of the bloodstain (in mm^"1) at wavelength λ; d ost is the optical pathlength through the host (in mm); and

dbi is the optical pathlength through the bloodstain (in mm).

Once the absorbance spectra of the bloodstain and the host material are determined, they can be used, together with the known spectra of blood, oxy- haemoglobin, met-haemoglobin and water as the input matrix for the fitting procedure to determine the relative concentrations of met-haemoglobin and oxy- haemog!obin in the bloodstain (40). This process is based on the Kubelka Munk Reflection Theory, which is a simplified solution of the radiation transport equation. This theory is employed to analyze the recorded diffuse reflectance spectra. The following formula was used:

Here / denotes the bloodstain's reflectance, and l₀ is the reflectance of the substrate surface. and S represent the absorption and scattering of the bloodstain. Here Κ=μ₃/η with μ₃ the absorption coefficient per unit length which depends on wavelength and η being a dimensionless function depending on albedo. For scattering S, Lorentz-Mie scattering was assumed:

S— Sn (2) The scattering S depends on the wavelength λ in nm, λ₀ is 450 nm and the scattering coefficient at 450 nm, S₀ is 13.5 mm^'! The scattering was set to be constant over the total measured time period.

Then, the known spectra of oxy-haemoglobin, met-haemoglobin and haemichrome are fitted (41) using a linear constrained optimisation fitting algorithm to the corrected spectra for oxy-haemoglobin, met-haemoglobin and haemichrome. In this case a linear least squares (LLS) fit was used. The LLS fitting algorithm varies the amplitudes of the three absorption spectra, in order to find the combination of the three with a minimum of difference between the theory and the diffuse reflectance spectrum. The LLS fitting procedure yields the estimated haemoglobin fractions of HbC>2, met-Hb and HC.

For correct data analysis, it is important that there is a high correlation between the reflectance signal and the LLS fit. If the correlation is poor, over compensation by one of the compounds may occur, and the outcome of the fitting procedure becomes unreliable. To prevent this, a quality test between data and fit is utilized (42). Thus, only LLS fits with correlation coefficient R²>0.98 were accepted. Two typical diffuse reflectance spectra with corresponding LLS fits are shown in Figure 9, where Figure 9(a) shows a spectrum for a bloodstain that is ¼ day old and Figure 9(b) shows a spectrum for a bloodstain that is 63 days old. As can be seen the LLS fit is fairly accurate in both cases. Measurements may also be carried out at different positions on the bloodstain to see whether the bloodstain is homogeneous, or whether there is a variation within a stain and what the variation is.

Figures 10a and 10b show that the amount of Hb0₂ decreases over time, because of the oxidation into met-Hb. Figure 10a comprises data up to around 60 days whereas Figure 10b comprises data up to around 500 days. It appears that the amount of met-Hb reaches a maximum after two weeks; met-Hb, in its turn, denatures into HC. At all moments, a unique combination of these three haemoglobin compounds exists, which can be utilized for adequate age determination. For performing age estimation, the haemoglobin fractions determined from the validation set are matched (43) with the comparison ("standard) graph in Figure 10a or 10b. This can be done by eye or by an appropriate matching algorithm. The relative differences between the measured Hb0₂, met-Hb and HC fractions on one hand and the values in the "standard" graph are calculated and the best match is determined. The best match can be considered to be found at a minimum difference of this sum. This thus provides the estimated age of the bloodstain (44). Alternatively, each component can be matched to the comparison graph and the age determined. The ages so determined can then be averages to provide an age estimation. Other techniques of making the age estimation based on a comparison of the measured HbC>2, met-Hb and HC fractions on one hand and the values in the "standard" graph can be used, as will be well known to a person skilled in the art.

Figure 11 shows the age estimation following the above described procedure for test samples where the age of the bloodstains was known to compare the estimated age with the known actual age. The combination of biological variation and uncertainties in the measurement procedure cause the spreading of the age estimation. For instance, at an age of seven days, the estimated age varies between four and ten days (±50). When measured at 55 days, the estimated age varies between 45 and 70 days (±25%r).

It will, of course, be appreciated that although the technique described above used reflectance spectroscopy, any appropriate type of spectroscopy, such as Raman or reflectance spectroscopy could be used.

A second embodiment of the process described above is shown schematically in Figure 12. In this case, the steps of warming up the lamp (34), determining the standard spectra for Hb0₂, met-Hb and HC (35), measuring the reflection spectrum for the host material (36) and measuring the reflectance spectrum for the bloodstain (37) are the same as in the process described above. The latter measurement of the spectrum of the bloodstain (which may be measured at different positions on the bloodstain) is, however, then repeated several or many times over the course of minutes, hours or even days, as is indicated by the arrow 45. If the measurements are taken at different positions on the bloodstain, then the measurements at different times will, of course, be taken at the same corresponding positions. The measurements are repeated at different times in order to determine the reaction rate, i.e. the change in the concentration of each of the components of haemoglobin. Thus, in this case, rather than determining the concentrations of each of the components, the rates of change of the concentrations are determined (46) after the absorbances of th e host materia! and the bloodstain have been determined (39) and the reflectance spectra have been corrected (40), as described above. Again, the determination of the rates of change of the concentrations is carried out using a LLS fit and a quality test of the fit is carried out (47). Thereafter, the rates of change of the concentrations, are compared (48) to the rates of change of concentrations of Hb0₂, met-Hb and HC, i.e. the slopes of the lines in the graph of Figures 10a, b. The position where the sum of the differences is at a minimum is considered to be the best match and the age of the bloodstain is determined (49).

As mentioned above with respect to Figures 10a, b the amount of Hb0₂ decreases over time, because of the oxidation into met-Hb. However, it will be appreciated that the rate of the reaction can vary depending on environmental conditions, with a number of environmental factors, for example, ambient temperature, humidity, thermal conductivity of a substrate on which the body or body sample is located, absorption by a substrate on which the body or body sample is located and/or other factors affecting the rate. Thus, it is envisaged that "standard" graphs can be constructed for a number of different combinations of environmental factors, for different typical environmental conditions. For example, there may be several possible "standard" graphs of the amounts of the fractions, for different environments, each graph being a combination of the effects of different environmental factors for a particular "standard" environment, such as outdoors day or night, air conditioned indoors, non-air conditioned indoors day or night, etc. These "standard" graphs can be determined experimentally, for example, as shown in Figure 13, which shows the experimental results from controlled experiments for the Hb0₂ fraction at 40% humidity at ambient temperatures of -20° C, 4° C, 23° C and 36° C. As can be seen, as the ambient temperature increases, the amount of HbG₂ present decreases more rapidly over time to a lower amount than at lower ambient temperatures. Such experimental graphs can be used to produce "standard graphs", by providing best fit lines to the experimental data, as shown in Figure 14.

Similarly, Figure 15 shows how ambient humidity affects the decrease in the amount of Hb0₂ present. The graph shows schematically the rates of decrease of the amount of Hb0₂ present at ambient humidities of 70%, 80% and 90% humidity. As can be seen, as the ambient humidity decreases, the amount of Hb02 present decreases more rapidly over time to a lower amount than at higher ambient humidities. Similar graphs can be constructed for other environmental factors, such as thermal conductivity of a substrate on which the body or body sample is located, absorption by a substrate on which the body or body sample is located and/or other factors affecting the rate, as well as graphs showing the amounts of the other components over time for the various environmental factors.

In order to carry out the comparison of the measured HbC½, met-Hb and HC fractions to the "standard" graph therefore, a better estimate can be obtained by first choosing the best "standard" graph based on an estimate of the environment where the sample was found at the estimated time that the bloodstain originated. By thus choosing the most likely "standard" graph for that environment, a better estimate of the age of the bloodstain can be achieved.

Of course, it will be appreciated, that each of the graphs can be considered to have an equation associated with it, with different parameters for the rates of change of the component fractions depending on the particular environmental factors. For example, for the graph of Figure 14, the various experimentally derived "standard" lines shown can be "converted" into an equation of the form:

Fraction HbO_t = 1 - A - Mtims - B)

where the parameters A and B take different values to provide the appropriate shape of the line at different temperatures. Of course, different "standard" graphs for different environmental factors may have different equations, not just different values of parameters A and B. Nevertheless, it will be apparent that, once appropriate experimental data has been gathered for the fractions of the components for different combinations of environmental factors, corresponding to different environmental conditions, then appropriate equations can be determined for each such graph.

These equations with different parameters can be stored in a database together with particular environment scenarios, each being associated with different set values of the various parameters. Thus, for example, an outdoor nighttime winter environment may have associated with it a lower temperature and a higher humidity than an outdoor daytime summer environment. It will be appreciated that many different such environmental scenarios can be stored, together with associated different values of the parameters of the equations, so that for any particular chosen environmental scenario, a particular set of associated values of the various parameters can be found from the database in order to determine the equation that produces the "standard" graph for that particular component fraction in that environment. Figure 16, shows a table of various values of the parameters A and B for the equation:

Fraction bO, = 1 - A^■ itime^■ B)

for the fraction of Hb0₂ for the graph shown in Figure 14. The database would normally be stored in computer memory, coupled to a computer processor. As can be seen, the database can include particular examples of environmental conditions, so that, if a particular environmental condition is chosen by a user on the computer, using a computer user interface, the particular environmental factors and the values of the parameters corresponding to that chosen environmental condition, are automatically read by the computer processor from the database and, if desired, inserted into the appropriate equation, with the "standard" graph then being generated from that equation by the computer. The measured values can then be compared to that graph, either visually, using the graph itself, or by analyzing the measured amounts by the computer processor to compare them to the generated equation. For example, if a bloodstain is found outside in summer and it is estimated that it was outside all night, then the Outside Summer Night environment may be chosen. This equates to a temperature of 4 °C and a humidity of 40%, and the computer would then find from the database that A= 0.0498 and b=3.8507 and generate the "standard" graph. If the measured Hb0₂ fraction is 0.7, then it can be determined, either visually from the generated "standard" graph, or by the computer from the equation without the need for generating the "standard" graph, that the estimated age will be 07 hours.

Of course, it will be appreciated that such an analysis could, in fact, be undertaken for a number of differing environmental scenarios to find the best fit and/or for different component fractions, as described above. For example, Figure 17 shows the experimentally determined graph for the HC component at a humidity of 40% for the same temperatures as those for which the Hb0₂ fraction was determined. Thus, the measured HC component could also be compared to a "standard" graph and the age determined from that, thus enabling a more accurate overall age determination to be made.

It will be appreciated that although only a few particular embodiments of the invention has been described in detail, various modifications and improvements can be made by a person skilled in the art without departing from the scope of the present invention.

For example, when a sample such as a blood stain is deposited on a dark or coloured background that absorbs much of the visible light then reflectance spectroscopy can have shortcomings for the identification and age estimation of the sample. The inventors have found that near infra-red reflectance spectroscopy can distinguish blood samples from other substances on white and coloured backgrounds.

Figure 18 shows white and coloured cotton t-shirt background spectra in the infrared range of the spectrum. For ease of comparison with literature data, the reflectance spectra is transformed to apparent absorbance spectra (log

1/reflectance or log1/R). We can see that coloured cotton only differs in the visible part of the spectrum less than 850nm but it the near infrared part of the spectrum all cotton colours have the same spectral features.

Example:

Blood was drawn from two healthy female volunteers and deposited on several backgrounds; brown cork and white, back, red, green and blue cotton. Next to three blood samples on each background, a reference sample of each plain background was created. Between measurements, all samples were stored in a laboratory with a stable room temperature and humidity.

It is known that variation between individual log (1/R) spectra is not only caused by the chemical composition of the sample, but also by different particle sizes, non-uniform light scattering and variable spectral path lengths. To minimize the spectral variation caused by non-chemical properties, several preprocessing methods (each described below) were applied to the log (1/R) spectra.

Preprocessing method Description

Offset correction Each spectrum is corrected by subtracting either its value at one wavelength or the mean value in a selected range. Multiplicative Scatter Each spectrum is shifted and rotated so that it

Correction (MSC) fits as closely as possible to the mean spectrum of the data.

Standard Normal Variate The mean of each spectrum is subtracted,

transformation (SNV) followed by division of the spectrum by its

standard deviation.

First derivative Differentiation by using the Savitsky-Golay

algorithm.

Second derivative Differentiation by using the Savitsky-Golay

algorithm.

For blood stain identification, a preferred preprocessing method would remove spectral variation between blood samples while conserving spectral variation between blood samples and other substances. We calculated the coefficient of variation at each wavelength for all preprocessed spectra, to determine which (preprocessed) spectral features are stable and thus suitable to use for the identification of blood stains.

For the purpose of age estimation, a preferred preprocessing method would remove spectral variation between blood samples of the same age while conserving spectral variation between blood samples of different ages. We calculated the coefficient of determination R² between the logarithm of the age and all preprocessed spectra at each wavelength. All coefficients of determination were compared to choose the most useful preprocessing method and to see which spectral changes correlate highly with the age of blood stains.

Partial Least Squares regression (PLS) was used to create a model predicting the age of blood stains based on the log (1/R) spectra. PLS is a useful statistical tool for the analysis of spectroscopic data, as it can handle datasets in which there are more variables than observations, and the data may contain highly correlated predictor variables. PLS makes linear combinations of the original predictor variables to construct new predictor variables, which are the most relevant for predicting the age. All ages of blood stains on white cotton and brown cork were estimated using a leave-one-out cross-validation (CV) and the root mean squared errors of cross validation (RMSECV) were calculated. Results were compared for models using the visible (Vis) wavelength range, the near infrared (NIR) range and a combination of these (Vis-NIR). The log (1/R) spectra of the differently coloured cotton backgrounds were compared visually. Based on this comparison, a wavelength region was chosen in which the colour of the background did not influence the spectra. This region was used to estimate the age of blood stains on coloured backgrounds, using a PIS model which was trained with the blood stains on the other backgrounds. Finally, the root mean squared errors of prediction (RMSEP) were calculated for all colours.

The highest coefficients of determination R² were found when using the first derivative spectra. In two regions of the first derivative spectra R^z was higher than 0.975. These regions are depicted in Figure 19. The first region (from 754 to 774 nm) was near a broad absorption peak of oxyhemoglobin. The second region (from 1700 to 1732 nm) was between two typical absorption peaks (at 1690 and 1740 nm) known to be present in the spectra of haemoglobin, albumin, globulin, and glucose. Figure 19 shows absorbance spectra of fresh (5 hours, solid line) and old (77 days, hatched line) blood stains corrected for a cotton background.

Figure 20 shows an age estimation on black cotton using the predetermined regions in the near infrared part of the spectrum. This mode! can be used to predict the age of the blood stains on the remaining background. The root mean squared error of prediction (RMSEP) was 8.9 %.

Claims

A method of dating a body sample comprising:

taking at least one spectroscopic measurement of the sample, the measurement including at least two predetermined positions in the spectrum, the positions having spectral characteristics corresponding to at least two predetermined substances present in the sample that have a time varying relationship with each other;

determining a measured relative concentration of each of the predetermined substances present in the sample from the spectroscopic measurement;

comparing the measured relative concentrations of the at least two predetermined substances with a known variation of the relative concentrations of the at least two predetermined substances over time; and

determining a good fit of the measured relative concentrations to the known variation of the relative concentrations, so as to provide an indication of the age of the sample.

A method of dating a body sample, comprising:

taking at least two spectroscopic measurements of the sample at different times, the measurements including at least two predetermined positions in the spectrum, the positions having spectral characteristics corresponding to at least two predetermined substances present in the sample that have a time varying relationship with each other;

determining a measurement of change in the relative concentration of each of the predetermined substances present in the sample from the spectroscopic measurements;

comparing the measurements of change in the relative concentration of each of the predetermined substances with a known variation of the changes in the relative concentrations of the predetermined substances over time; and determining a good fit of the measurements of changes in the relative concentration to the known variation of the changes in the relative concentrations, so as to provide an indication of the age of the sample.

A method of dating a body sample according to claim 2, wherein, if the changes in the relative concentrations of each of the predetermined substances are small, the method further comprising the method of claim 1.

A method of dating a body sample according to any one of claims 1 , 2 or 3, wherein the sample is a body fluid.

A method of dating a body sample according to claim 4, wherein the body fluid is blood.

A method of dating a body sample according to claim 5, wherein the predetermined substances comprise two or more of oxy-haemoglobin, met-haemoglobin and haemichrome.

A method of dating a body sample according to claim 6, wherein the measurement(s) includes at least three predetermined positions in the spectrum, the positions having spectral characteristics corresponding to at least three predetermined substances, the three predetermined substances comprising oxy-haemoglobin, met-haemoglobin and haemichrome.

A method of dating a body sample according to any preceding claim, wherein the spectroscopic measurements) comprise reflectance, Raman or fluorescence spectroscopic measurement(s).

A method of dating a body or body sample, the method comprising:

measuring one or more time-varying parameters of the body or body sample, the time-varying parameter(s) having a rate of time variance that changes according to at least one environmental factor; provsding a database of rates of time variance of the time-varying parameter(s) for the at least one environmental factor;

estimating the at least one environmental factor for the particular environment(s) where the body or body sample was located prior to the at least one measurement being made; and

determining an estimate of the age of the body or body sample by utilising the at least one measurement of the or each time-varying parameter and the rate of time variance for that time-varying parameter for the estimated at least one environmental factor.

10. A method of dating a body or body sample according to claim 9, wherein the time-varying parameter comprises temperature.

1 1. A method of dating a body or body sample according to claim 9, wherein the time-varying parameter(s) comprises two or more of:

concentration of oxy-haemoglobin;

concentration of met-haemoglobin; and

concentration of haemichrome. 12. A method of dating a body or body sample according to any one of claims 9, 10 or 11 , wherein the environmental factor comprises one or more taken from:

ambient temperature;

ambient humidity;

absorption by a substrate on which the body or body sample is located. 13. An apparatus for dating a body sample comprising:

a device for taking at least one spectroscopic measurement of the sample, the measurement including at least two predetermined positions in the spectrum, the positions having spectral characteristics corresponding to at least two predetermined substances present in the sample that have a time varying relationship with each other; and

a processing device for determining a measured relative concentration of each of the predetermined substances present in the sample from the spectroscopic measurement, comparing the measured relative concentrations of the at least two predetermined substances with a known variation of the relative concentrations of the at least two predetermined substances over time, and determining a good fit of the measured relative concentrations to the known variation of the relative concentrations so as to provide an indication of the age of the sample.

An apparatus for dating a body sample comprising:

a device for taking at least two spectroscopic measurements of the sample at different times, the measurements including at least two predetermined positions in the spectrum, the positions having spectral characteristics corresponding to at least two predetermined substances present in the sample that have a time varying relationship with each other; and

a processing device for determining a measurement of change in the relative concentration of each of the predetermined substances present in the sample from the spectroscopic measurements, comparing the measurements of change in the relative concentration of each of the predetermined substances with a known variation of the changes in the relative concentrations of the predetermined substances over time, and determining a good fit of the measurements of changes in the relative concentration to the known variation of the changes in the relative concentrations so as to provide an indication of the age of the sample.

15. An apparatus for dating a body sample according to claim 14, wherein, if the changes in the relative concentrations of each of the predetermined substances are small, the processing device further determining a measured relative concentration of each of the predetermined substances present in the sample from the spectroscopic measurement, comparing the measured relative concentrations of the at least two predetermined substances with a known variation of the relative concentrations of the at least two predetermined substances over time, and determining a good fit of the measured relative concentrations to the known variation of the relative concentrations so as to provide an indication of the age of the sample.

16. An apparatus for dating a body sample according to any one of claims 13, 14 or 15, wherein the sample is a body fluid. 17. An apparatus for dating a body sample according to claim 16, wherein the body fluid is blood. 8. An apparatus for dating a body sample according to claim 17, wherein the predetermined substances comprise two or more of oxy- haemoglobin, met-haemoglobin and haemichrome.

19. An apparatus for dating a body sample according to claim 18, wherein the measurement(s) includes at least three predetermined positions in the spectrum, the positions having spectral characteristics corresponding to at least three predetermined substances, the three predetermined substances comprising oxy-haemoglobin, met-haemoglobin and haemichrome.

20. An apparatus for dating a body sample according to any one of claims 13 to 19, wherein the spectroscopic measurement(s) comprise reflectance,

Raman or fluorescence spectroscopic measurement(s).

21. An apparatus for dating a body or body sample, comprising:

a device for taking at least one measurement of one or more time- varying parameters of the body or body sample, the time-varying parameter(s) having a rate of time variance that changes according to at least one environmental factor; and

a processing device for obtaining rates of time variance of the time-varying parameter(s) for the at least one environmental factor, for the particular environment(s) where the body or body sample was located prior to the at least one measurement being made, from a database of different rates of time variance of the time-varying parameter(s) for different environments, and determining an estimate of the age of the body or body sample by utilising the at least one measurement of the or each time-varying parameter and the rate of time variance for that time- varying parameter.

An apparatus for dating a body or body sample according to claim 21 , wherein the time-varying parameter comprises temperature.

An apparatus for dating a body or body sample according to claim 21 , wherein the time-varying parameter(s) comprises two or more of:

concentration of oxy-haemoglobin;

concentration of met-haemoglobin; and

concentration of haemichrome.

An apparatus for dating a body or body sample according to any one of claims 21 , 22 or 23, wherein the environmental factor comprises one or more taken from:

ambient temperature;

ambient humidity:

absorption by a substrate on which the body or body sample is located.